In one of the chapters in Prof. Giulio Tononi’s book “Phi: A Voyage from the Brain to the Soul,” the hero faces a contraption reminiscent of Kafka’s pain machine from “In the Penal Colony.” Here, too, a wretched condemned man lies inside the machine, while an enthusiastic officer presents its capabilities to a chance spectator who expresses interest in what is happening. But whereas Kafka describes, with bloodcurdling precision, how the machine’s vibrating needles slice through the skin of the man condemned to death and inscribe the indictment on his back, Tononi offers a more sophisticated machine, whose needles do not even touch the condemned man’s body yet generate inconceivable pain.
Tononi’s needles embroider and unravel, with chilling precision, connections between the neurons in the exposed brain of the condemned. They do so not as punishment for the soldier’s insubordinate conduct, but rather as part of the project of an ambitious officer who wants to study the repertoire of human pain. “They want crude pain,” he complains of his superiors, “common pain, pain like gravel.” They seek confession to a crime, whereas he, artist that he is, deals in creative expression: The invention of new nuances of pain, the design of pure suffering, the creation of torments that do not pass through the simple tortures of the body, but rather are delivered directly from the mind by stimulating it alone.
This grueling scene, described with great talent, is but one of many in the book, which portrays what several scientists have termed the most promising theory of consciousness today. Tononi, who also developed the theory, deals with one of the most problematic issues in the mind-body field: the manner in which we experience things. What exactly does a person feel when he experiences pain, tastes wine or sees the orange color of the sunset? Can one person’s experience of pain be compared to that of another? And what causes pain to be experienced as such and not as darkness, for example, or the smell of rain, or shame? Philosophers use the term “qualia” (“quality of consciousness”) to describe this mental aspect, and many of the best have tried their hand at resolving this issue within the framework of what has been dubbed the mind-body problem.
But Tononi is not a philosopher. As a psychiatrist and neuroscientist at the University of Wisconsin-Madison in the United States, he attempts to answer the question in a reductionist fashion − in other words, to explore the concept of consciousness as a product of the brain’s neurons and the chemical substances that connect them.
As a writer, Tononi presents the theory in an original way that combines reference work and fiction, historical reality and fairy tales. Tononi makes use of prose, poetry, and the plastic arts to explain his arguments, which gives the reader an uplifting reading experience − if not necessarily proof of the theory’s scientific validity.
The book’s protagonist is Galileo Galilei, who we encounter lying in his bed, dreaming. At first he observes the sky and afterward the surface of his own brain, and tries to discover the rules of the latter as he did those of the former. Later on, he walks through the corridors of an abandoned building − a hospital or monastery perhaps − and encounters an array of characters and situations that shed light on the essence of consciousness and its locus in the brain.
In a telephone conversation with Tononi, he explains why he chose this unique literary genre. “Consciousness,” he says, “is one of the few subjects to which both the natural sciences and the humanities assign great significance. It is a link that connects the two disciplines, so I thought it would be a challenge to connect the fields not only on the level of content, but also in terms of style. I wanted to create something that most people could understand, or at least connect with to one degree or another, without getting into the details of the science and mathematics relating to the theory.”
Why did you choose Galileo as your protagonist?
“It was a natural choice. Galileo is considered the first scientist ever because he removed the observer from nature, and thereby pioneered the path to the objectivity of science. He said let’s not talk about the subjective things that depend on the personal experiences of the researcher, but rather concentrate on the objective things. We’ll deal with the motion of bodies and not with the experience of the researcher when he travels; we’ll study the celestial objects and not the sense of wonder they arouse.
“By contrast, when we study consciousness, we are dealing precisely with the subjective experience of man. And so I thought there is no one more worthy than Galileo of being the one to restore the spectator, the researcher, to the nature we are trying to understand.”
Is that even possible? The prevailing philosophical approach is that this is an unsolvable problem in principle. The claim is that subjective mental processes cannot be studied because they are unique to a particular person and also private, so nobody but he can grasp them. This stands in contrast to physical processes that are objective and “public” and hence open to investigation, measurement and comparison.
“I think it is absolutely possible to study the phenomenon. My theory takes a scientific approach to man’s unique, subjective experience. And as I claim in the book, every experience can be treated in mathematical terms and even evaluated as a number.”
Tononi’s theory is called integrated information theory, and as a first step toward understanding it, he begins with a lovely thought experiment in which Galileo takes part, a hint at the thought experiments the real, historic Galileo frequently conducted. The fictional Galileo enters a room on the orders of his escort, and sits in a chair. Because the room is empty and its walls are white and completely bare, Galileo looks about him and does not notice anything in particular. The escort gives him a simple assignment: To say “light” or “dark” according to the state of the lighting in the room. He is told not to dawdle and not to allow his thoughts to wander, but rather to concentrate and make do with the plain report. After the briefing, the light in the room goes off, and immediately comes on again. Then darkness reigns again and then light appears once more, and so on and so forth. Galileo responds promptly. He says “light” one time and “dark” another, depending on the situation. Not an overly complicated task, he thinks, once everything is over.
The escort points at the chair Galileo is sitting on. Beneath it is a photodiode − an astonishingly simple device, perhaps the simplest ever built, a type of light detector that switches on by itself when hit by light, and turns off when it goes dark. And thus, when Galileo reports whether it is light or dark, the photodiode is doing exactly the same thing: turning on and off, accordingly.
To an outside observer, Galileo’s response appears identical to that of the photodiode, and they are indistinguishable. Nonetheless, it is clear to us that the photodiode does not possess consciousness in relation to the light and the darkness, whereas Galileo does. Galileo has a unique subjective sensation when he experiences darkness, whereas the photodiode does not experience a thing.
What is there in Galileo, and not in the photodiode, that creates consciousness in the former?
“In contrast to the photodiode, Galileo’s mind does not distinguish only between two states − darkness and light − but rather between darkness and countless other states. For Galileo, darkness is not merely not-light. It is also not-red or any other color, not-table or any other piece of furniture, not-Rome, not the face of his beloved, not-Eucalyptus scent and not a billion other things. For the photodiode, on the other hand, darkness is only one state out of two.”
In other words, in Galileo’s mind an enormous repertoire of possible states exists, and you claim consciousness stems from that.
“Yes, choosing one state out of a huge variety of states is a condition for the existence of consciousness. It has to do with the definition of information as Claude Shannon formulated it in the middle of the last century. Shannon was a U.S. mathematician and electrical engineer who developed the information theory that pioneered the way for the field of computers and telecommunication. Shannon argued that information is the removal of uncertainty. Information enables us to know something with greater certainty compared to the level of certainty we had before we received the information. The information can be measured and evaluated quantitatively, by checking how much it lowered the level of uncertainty in the system. Shannon showed this mathematically, but it can also be understood intuitively.
“A lot of information is when, to begin with, there are a lot of options, and the particular information you received narrows them down greatly: there was great uncertainty and now there is much less. Little information is when there isn’t much difference between the uncertainty you had before the information and after it.”
As happens with weather forecasts. When the weatherman tells us in the summer that the next day is going to be hot, there isn’t much information because that is the only option there is. On the other hand, if he says that in winter, he is giving us a lot of information, because in winter − the Israeli one at least − there is a lot of uncertainty: The weather can be very cold, cold, pleasant or hot. In winter, the weather forecast contains a lot of information because it decides among numerous possibilities, which is to say, reduces a lot of uncertainty.
“Exactly so. Therefore, if we go back to Galileo and his photodiode, we will see that a switched-off photodiode is a state in which there is little information because it removes only a little bit of uncertainty: one possibility out of two. On the other hand, when Galileo’s mind recognizes darkness, he is in a state of a lot of information because it is one possibility that was determined from among billions – in other words, removal of plenty of uncertainty.”
Still, consciousness is not merely a lot of information. After all, if we were to just hook up a lot of photodiodes simultaneously, we would create a system that contains a lot of information but we would not create consciousness.
“Right. Consciousness is what I call ‘integrated information’: information that is a whole greater than the sum of its parts. Take, for example, a digital camera. It has a board with a million photodiodes, each of which can be in one of two states. Which is to say, the camera board is very rich in information, and nonetheless it is clear that the camera has no consciousness because each part of it, each photodiode, operates independently of the other. If we take the photodiode board and cut it into two, four or eight parts, nothing will change in terms of the information stored in the camera. The photographed picture will remain exactly the same before and after the cut.
“In the brain, on the other hand, the neurons ‘talk’ to each other, intercommunicate and integrate information from different parts of the brain to create a whole that is greater than the sum of its parts.
“You will never be able to perceive the shape of an object without at the same time also perceiving its color, size and orientation. All the information comes to you at one and the same time because of the intersections of information that pass between many different centers of neurons. The name of the game is connectivity, and this integrated information − the whole greater than the sum of its parts, which can’t be broken down into its parts – is consciousness.”
Philosophers such as Immanuel Kant discussed the notion that every experience is unified, and cannot be otherwise. They have already argued that you cannot break experience down into its parts i.e., separate seeing from hearing, for example, and grasp them as two separate entities. What is new about your theory?
“The concept which holds that consciousness is a unity indeed has a long philosophical history. But what had not been realized is that it is the combination of unity and diversity − integration and differentiation − that is the hallmark of consciousness. The integrated information theory says that for consciousness to be present, the brain must work like a single entity − this is the integration part. But at the same time, the brain must have at its disposal a vast repertoire of different states, so that when it is in a particular state, it means what it means because it differs in its particular way from all the others − that is the information part.
And indeed, here’s an interesting fact that goes along with the theory. The cerebellum has far more neurons than the cerebrum − four times as many − but they are not interconnected, so the cerebellum is not integrated, and indeed the cerebellum has no consciousness. This is why injuries to the cerebellum will never cause a loss of consciousness. By contrast, the cerebrum, while having fewer neurons, is wonderfully integrated. And indeed, injuries to the cerebrum can definitely lead to a loss of consciousness or to its splitting, because they violate the connectivity between the neurons.”
When does consciousness get lost?
“It happens in death, of course, but also in people who are anesthetized for the purpose of surgery and in people defined as being in a vegetative state. And besides that, it happens to all of us nightly, and, if you will, that is one of the ways of defining consciousness: that thing that dissipates when we dive into a deep, dreamless sleep. In that state, our entire private universe disappears − people and things, colors and sounds, pleasures and pains, thoughts and feelings. Our sense of self disappears too, and when we are woken up at this stage, we feel like we’ve emerged from some experience-free void. In my opinion, the sentence ‘I think, therefore I am’ should have been phrased ‘I have consciousness, therefore I am,’ because consciousness is the only thing we are certain exists.
“The brain is a kaleidoscope of neural states,” says Tononi, “that are constantly changing and redesigned by each experience. The experience of darkness, for example – ‘the qualia of darkness’ as the philosophers would call it – occurs when certain nerve fibers and not others ‘switch on’ together in different centers of the brain. These ‘lit-up’ fibers constitute a unique state, which is different from the state that is formed by another experience. The perception ‘this is a chair’ constitutes a template of brain activity, which is different from the template that is formed by the experience of ‘the taste of the wine’ or by the experience of ‘the pain.’”
So, according to your theory, there is no particular point where consciousness resides. There is no center that sums up all the sources of information into one overall picture.
“There is no one who sees everything and then makes decisions. Consciousness is embodied in the system and is not dependent on an outside observer.”
What is self-awareness? That observer who watches what we see, hear, smell, and remembers, and expresses his opinion about them?
“Self-awareness is part of consciousness, but it is not consciousness as a whole. Consciousness is the subject in general: my unique, subjective perception of the world. My private experience. The sense of self within that experience is just another aspect of consciousness.”
Attributing consciousness to the complexity of the connections in the brain suggests that consciousness developed through evolution. Is it possible to point to a particular stage at which consciousness crystallized into an entity that distinguishes between us and other animals?
“It seems to me safe to say that apes have a certain kind of consciousness, but that doesn’t take integrated information theory. They are so similar to us, after all, that clearly they have consciousness of a sort. Precisely when in evolution a consciousness close to what we have began is hard to determine, but there is no doubt that consciousness is part of an animal’s intelligence and a capacity that has high survival value. The ability to integrate information enables us to get meanings out of diverse context, and thereby cope promptly with new situations we did not encounter previously. If, for example, it suddenly grows dark at noon, the context can suggest to us whether we are dealing with a big cloud that has covered the sun or with an Iranian bomb.”
With a materialist attitude like that, it stands to reason you believe that consciousness can be accorded to machines.
“I see no reason why not. One of the classic thought experiments suggests exchanging neurons for silicone chips, one cell at a time, and thereby create consciousness that is not dependent on material, but rather on the information embodied in it. So long as the chips or artificial neurons generate the same three-dimensional shape of the qualia, the very same experience will be created. Naturally, it would be possible this way to create entirely new sensations.”
In a witty ars poetica reference, you write at the end of one chapter: “Perhaps the author was trying to avoid equations at all costs, but the result is far from satisfactory.” The thing is, a confession, even if graceful and sophisticated, does not grant its maker absolution and does not exempt him from presenting scientific proof of his theory’s validity.
“As the author of a popular science book, I wanted to refrain from bringing experiments and presenting mathematical equations to make it easier to read. You’ll have to believe me that integrated information theory has a mathematical formulation, which is based on using branches of mathematics such as information theory and graph theory.
“As I explain in the book, we have defined a value, whose symbol is the Greek letter phi, and it expresses the level of integrated information that is in a system.
“Phi depends on the system’s degree of connectivity: when it is low, i.e., when there is not much information sharing among the parts, its value is low. That, for example, is what happens in the cerebellum. When connectivity is high, and different parts of the system do different things, then the value of phi is high, and that happens in the cerebrum. The cerebrum is made up of several specialized centers − sight, hearing, speech, etc. − but they all talk to each other.
“Phi expresses the synergy of the system, meaning it measures how much greater the information of the whole system is than the sum of information of its parts. Right now we are calculating the phi in very simple systems, which have only a limited number of parts that communicate with each other. But the idea is to apply this mathematics to complex systems as well.”
You ascribe to consciousness a mathematical value that can be calculated, a number. In our perception, consciousness has meaning − my pain does not have the same meaning for me as your pain does for you. Darkness, love, longing have unique and personal meanings to each individual. How can these be reconciled?
“That is precisely what Shannon understood about information: that you can ascribe a numerical value to it, regardless of its meaning. Contrary to prevalent thinking before him, he ignored the meaning of the information and began viewing the information as mathematical magnitude. That was his greatness, and he thereby paved the way for the revolution in communication and data storage. This is also what is at the bottom of my addressing consciousness as information − the ability to treat it as mathematical magnitude.”
Let’s say your mathematical model can in fact describe consciousness. It is still a formal phrasing of a theory that also has to contend with biological reality. In other words, the theory is supposed to provide an explanation for various physiological phenomena, stand the test of experiments, allow for prediction of states, and be refutable. Does your theory meet these characteristics?
“I’ll tell you, for example, about a series of experiments we are doing now to test the brain in various states of consciousness. As you know, there are several stages of sleep, which appear in a set order: light sleep, dreaming sleep and deep sleep. Each of the stages is characterized by a different state of consciousness, but what is interesting is that when you measure the activity of the nerve cells in the brain, it is fairly similar at all the stages. The difference comes to light when you look not at the lone neurons but at the entire system and, more precisely, at its connectivity. Then you discover a major difference between the deep, unconscious, sleep and the other states, where there is consciousness. In a state of ‘ordinary,’ not deep, sleep, the neurons in the cerebral cortex talk to each other and communicate energetically, almost like when the brain is awake. In the deep-sleep stage, on the other hand, which is characterized by unconsciousness, the level of connectivity plummets. This behavior of the brain in the various stages of sleep supports the theory, which conditions the existence of consciousness on connectivity between the centers of information; on integrated information.
“When we monitor what is called ‘brain waves,’ we see that in the stages of
ordinary sleep, the brain is like a sea whose shifting movement is created by little waves moving in every direction. By contrast, when sleep deepens the waves become big and slow, and start working synchronically. The theory claims that this uniformity is what causes the disappearance of consciousness, because it entails the disruption of communication in various directions between the nerve centers. The uniformity does not permit information to pass between the parts of the brain, meaning that it does not generate integrated information, and so it dismantles consciousness.
“To test this claim, we use a powerful magnet (transcranial magnetic stimulation) that is attached to the scalp. We send the brain localized stimulation and monitor the brain’s response to it in people who are in various states of consciousness. What we are finding matches precisely what is expected according to the theory. When the stimulation is done to the brain of a wakeful person or the brain of one who is dreaming − both states that are characterized by high consciousness − we get a response of several nerve centers working together. The brain reacts as a single coordinated entity.
“On the other hand, when that same person enters a state of deep sleep − about 40 minutes after falling asleep − and we employ the very same stimulation on him, something completely different happens. The region of stimulation registers a similar response to the one that registered during wakefulness and dreaming, but in the other parts of the brain there is no response. It seems as though the brain lost its connectivity, so that instead of one complex entity, we get parts that function as isolated islands.
“It is fascinating, because the physical connections between the neurons in the various centers still exist, even in a state of deep sleep. It’s only the integration of information between them that disappears. It is dismantled consciousness, and now we are trying to understand what the mechanisms are that actually disconnect the neural connections. We also did this experiment with people under anesthesia and received the same picture − dismantling of the integrated information − and the same goes for people defined as being in a vegetative state.
Taking your definition of integrated information, we can find this not only in the brain, but also in many other complex systems: the digestive system, the iPhone, an ant colony. At any given moment, each of these systems is in one particular state out of an enormous range of possible states; each of them is a whole whose functioning is greater than the sum of its parts and cannot be separated into its parts without falling apart completely. Nevertheless, I imagine you would not define any system like that as possessing consciousness.
“No, for consciousness to be there, it is not enough for a system to be complicated, and to be made of interacting parts. It has to be made of many parts interacting in a very special way. It turns out that building a system that is both integrated and differentiated − one that is truly unified and yet has an immense repertoire of different states − is not easy at all. But somehow, that was the way our brain evolved. And we are all the more lucky for it − for good or bad. Otherwise, the universe would be truly a dark place − a place without the light of consciousness.”